Demodulating digital video broadcast signals

Abstract

Apparatus for demodulating digital video broadcast signals with an improved automatic frequency control comprises data modulated on a multiplicity of spaced carrier frequencies, including: analog to digital conversion device for providing a series of digital samples of the broadcast signal, Fourier Transform for analysing the samples to provide a series of data signal values for each carrier frequency signal processing devices for processing the series of data signal values including the phase-error-correcting, and automatic frequency control device for controlling the frequency of the signals input to the Fourier Transform Processor, wherein the automatic frequency control device includes coarse frequency control unit for controlling the frequency in terms of increments of the carrier spacing frequency, and fine frequency control unit for controlling the frequency for values less than a single carrier spacing frequency interval, wherein the coarse frequency unit includes recursive of filtering for assessing a predetermined number (N) of carrier signals on either side of the nominal position of a predetermined continual pilot signal, wherein the outputs of the filter are provided to respective memory locations of a memory, the memory being divided into a plurality of sections, first device for determining within each section a first signal best representing the continual pilot signal, and device means for determining from among the first signals from the plurality of sections, a second signal which best represents the continual pilot signal.

Description

This application claims benefit of provisional application Ser. No. 60/054,195 filed Jul. 30, 1997.

This invention relates to demodulating digital video broadcast (DVB) signals.

There are currently two major types of DVB, namely, terrestrial broadcasting and satellite/cable broadcasting. The invention is particularly, though not exclusively concerned with terrestrial broadcasting, which has special problems, particularly in communication channel impairment, arising from adjacent television channels, multipath and co-channel interference, for example. A type of transmission which has been developed to meet these problems is known as Coded Orthogonal Frequency Division Multiplexing (COFDM)-see for example "Explaining Some of the Magic of COFDM" Stott, J. H.--"Proceedings of 20th International Television Symposium, Montreux, June 1997. In COFDM, transmitted data is transmitted over a large number of carrier frequencies (1705 or 6817 for DVB), spaced (by the inverse of the active symbol period) so as to be orthogonal with each other; the data is convolutionally coded, to enable soft-decision (Viterbi) decoding. Metrics for COFDM include Channel State Information (CSI) which represents the degree of confidence in each channel for reliably transmitting data.

Modulation and Demodulation of the carriers may be carried out by a Fast Fourier Transform (FFT) algorithm performing Discrete Fourier Transform operations.

Subsequent to demodulation, signal processing corrections are carried out such as channel equalisation, channel state information correction, phase error correction, and automatic frequency control. The demodulated and corrected signal may then be decoded in an FEC (forward error correction decoder) for recovery of data.

Automatic frequency control (AFC) is important, since frequency offsets may appear after down-conversion to an intermediate frequency, because of variations in local oscillator frequency. Such frequency offsets are lethal for frequency recovery, and must therefore be reduced to a minimum.

In regard to phase error correction, a principal problem is that of local oscillator phase-noise. The addition of local oscillator phase noise to an COFDM signal has two notable effects;

1) To rotate the received constellation by an amount which is the same for all the carriers within any one OFDM symbol, but varying randomly from symbol to symbol. This is called the Common Phase Error (CPE), and primarily results from the lower-frequency components of the phase-noise spectrum; and

2) To add Inter-Carrier Interference (ICI) of a random character similar to additive thermal noise. ICI primarily results from the higher-frequency components of the phase-noise spectrum. ICI cannot be corrected and must be allowed for in the noise budget. It can be kept small in comparison with thermal noise by suitable local oscillator design.

GB-A-2307155 describes (see Sec 2.1, p.7 and FIG. 11) an arrangement for automatic frequency control wherein it is recognised that the phase of the retrieved signals in OFDM is proportional to frequency error, and therefore a signal representing phase can be used to control the frequency of a local oscillator.

In practice, the frequency offset arising from local oscillator frequency variations may extend over an interval of many carrier frequencies. One technique which has been used in other applications, using QPSK or QAM modulation, to recover such large frequency offsets, is frequency sweeping. Frequency sweeping involves sweeping over a first frequency range, to detect a frequency lock condition, and then waiting for some time to allow the frequency lock algorithm to converge for a reliable detection of frequency lock. If there is no convergence, then the sweep range is increased until lock is reliably detected.

Whilst this method is suitable for QPSK or QAM modulation because an AFC lock indication can be provided very quickly, a similar method applied to COFDM would take an unduly long time for convergence, because an OFDM symbol lasts 1 millisecond, which would be the basis for any frequency detection algorithm.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a method of automatic frequency control in a receiver for COFDM signals.

The invention is based on the recognition that, although the phase variation between adjacent symbols in COFDM is random, for continual pilot signals, as defined in the ETSI Specification, the intended phase of the signals is the same in adjacent symbol intervals. The phase error in adjacent data symbols may be determined by measuring the phase difference in adjacent symbol intervals in the continual pilot signals. Whilst this phase difference is primarily of use for common phase error correction, nevertheless it may also be employed for automatic frequency control since frequency variations are proportional to the change of phase.

Further, in accordance with the invention, it is recognised that frequency control may be split into two separate controls, namely coarse control for frequency offsets of integral numbers of carrier spacing intervals, and fine frequency control for frequency offsets of fractions of a carrier spacing interval.

For frequency offsets less than one carrier interval, the phase change may be used as representing the fine frequency offset. For coarse frequency control, a signal is used representing rate of change of phase. Since the phase variation between adjacent symbol intervals in continual pilots is constant, a second difference of phase error representing rate of change of phase error should be zero. This therefore provides a means of locating the continual pilot where the coarse frequency offset is a plurality of carrier spacings from the nominal position.

In our copending application (GBP12427A), there is claimed apparatus for demodulating digital video broadcast signals comprising data modulated on a multiplicity of spaced carrier frequencies, including:

analog to digital conversion means for providing a series of digital samples of the broadcast signal, transform means for analysing the samples to provide a series of data symbol values for each carrier frequency, signal processing means for processing the series of data signal values including phase error correcting means, and automatic frequency control means for controlling the frequency of the signals input to the transform means,

wherein the automatic frequency control means includes coarse frequency control means for controlling the frequency in terms of increments of the carrier spacing frequency, and fine frequency control means for controlling the frequency for values less than a single carrier spacing frequency interval,

wherein the coarse frequency means filter means for assessing a group of a predetermined number (N) of carrier signals on either side of the nominal position of a plurality of predetermined continual pilot signals to determine which signal best represents the continual pilot signal, whereby to determine the coarse frequency error.

Problems arise in that where large frequency offsets may be encountered, the number (2N+1) of carrier signals that needs to be assessed becomes large, and the nominal position of more than one continual pilot may be present in the search range. Hence a false minimum value may be assigned as the continual pilot being located. A further problem is that because of the wide search range, the overall filtering effect is diminished and the immunity to noise is decreased.

The present invention provides apparatus for demodulating digital video broadcast signals comprising data modulated on a multiplicity of spaced carrier frequencies, including:

analog to digital conversion means for providing a series of digital samples of the broadcast signal, transform means for analysing the samples to provide a series of data symbol values for each carrier frequency signal processing means for processing the series of data symbol values including phase error correcting means, and automatic frequency control means for controlling the frequency of the signals input to the transform means,

wherein the automatic frequency control means includes coarse frequency control means for controlling the frequency in terms of increments of the carrier spacing frequency, and fine frequency control means for controlling the frequency for values less than a single carrier spacing frequency interval,

wherein the coarse frequency means includes a filter means for assessing a group of a predetermined number (N) of carrier signals on either side of the nominal position of a plurality of predetermined continual pilot signals, wherein the output of the filter means for each carrier assessed is provided to respective memory locations of memory means, the memory means being divided into a plurality of sections, first means for determining within each section a first signal best representing the continual pilot signal, and second means for determining from among the first signals from the plurality of sections a second signal which best represents the continual pilot signal.

As preferred, the rate of change of phase of each of said predetermined number of carrier signals between consecutive symbol intervals is determined, and applied as inputs to respective filters of the bank of filters. Preferably, the filter means comprises a recursive filter for providing an output representative of the accumulated value of a plurality of previous input values, and the minimum output value of the output of the bank of filters is selected.

Various configurations of filter may be envisaged. It would be possible to provide a separate filter for each carrier signal or group of carrier signals assessed, which would however be very expensive in terms of hardware. However it is realised in accordance with the invention that the coarse frequency error will normally be the same for all carriers, and hence a filter can be arranged to operate on samples of carriers which are the same spacing from the theoretical position of the pilot in consecutive groups of carriers. Thus usually filtering will be carried out over a number of symbol intervals, and in each symbol interval a plurality of pilot signals are assessed. In accordance with the invention, it is preferred in order to reduce hardware to provide a single filter means with access to said memory means in which the filtered samples for each carrier position are stored and accessed and updated as required. As an alternative, a respective filter means may be provided for each of said memory sections, to process the carriers associated with the respective memory section.

BRIEF DESCRIPTION OF THE DRAWINGS

A preferred embodiment of the invention will now be described with reference to the accompanying drawings, in which:

FIG. 1 is a schematic block diagram of digital terrestrial front end converter incorporating the present invention;

FIG. 2 is more detailed block diagram of demodulating apparatus according to the invention forming part of the converter of FIG. 1;

FIG. 3 is a schematic view of a chip incorporating the apparatus of FIG. 2;

FIGS. 4A, 4B and 4C are diagrams illustrating the recovery of coarse frequency error;

FIG. 5 is a block diagram illustrating apparatus for recovering coarse frequency error;

FIG. 6 is a schematic waveform diagram illustrating a problem with recovery of coarse frequency error;

FIG. 7 is a block diagram of apparatus for recovering coarse frequency error in accordance with the invention;

FIG. 8 is a more detailed block diagram of a common phase error/AFC circuit together with a channel equaliser circuit;

FIG. 9 is a detailed block diagram of coarse AFC control, and FIG. 9A is an associated waveform diagram;

FIG. 10 is a block diagram of a circuit for providing CPE and AFC control from the circuits of FIG. 9; and

FIG. 11 is a table showing positions of continual pilot signals in the DVB-T 2K and 8K modes.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The preferred embodiment of the present invention comprises a front end for digital terrestrial television transmitted according to the DVB-T specification. The front end consists of two separate components. First, an analog down-converter that converts the input signal from UHF to a low IF. Second, an integrated circuit that accepts the analog signal from the down-converter and performs the required DSP operations, which include synchronisation and demodulation, to form a stream of soft decisions suitable for presentation to an FEC decoder (Forward Error Correction decoder).

Full compliance to the DVB-T specification means that the chip is capable of decoding signals transmitted in the following modes:

1) A signal that contains either 1705 or 6817 active carriers, commonly referred to as 2K and 8K respectively. The chip includes the functionality and memory required to perform the FFT algorithm in both modes.

2) Non-hierarchical QPSK, 16-QAM and 64-QAM constellations.

3) Hierarchical 16-QAM and 64-QAM constellations, either uniform or non-uniform with the possible scale factors α=2 and α=4.

4) Guard intervals 1/4, 1/8, 1/16 and 1/32 of the OFDM symbol length.

5) Viterbi code rates 1/2, 2/3, 3/4, 5/6 and 7/8.

Referring now to FIG. 1, a block diagram of the front end system, a down-converted 2 receives the input UHF broadcast signal from an antenna 4 and converts the carrier signal to a first IF frequency of 30-40 MHz and then to a second IF frequency of 4.57 MHz. Since the bandwidth of the modulated data is about 7.6 MHz, this second IF signal is sufficiently low in frequency to present the signal as a base band signal to a demodulator chip 6. Demodulator chip digitises the incoming signal at a rate determined by a voltage controlled oscillator 8, and provides an Automatic Frequency Control on a line 9 to down-converter 2. The output of demodulator 6 represents demodulated data and is fed to a FEC decoder 10 (Forward Error Correction or Viterbi decoder) for recovering the data. The decoded data is fed to a transport stream demultiplexer 12 and then to an audio-visual decoder 14. The front end is controlled by a system microcontroller 16.

Referring now to FIG. 2, this shows the demodulator chip 6 in more detail. The chip itself is shown schematically in FIG. 3. The low second IF is fed to an analog digital converter which samples the IF signal at a frequency of 18.29 MHz (×4 the second IF frequency of 4.57 MHz), in an analog to digital converter 20. The digital output samples are fed to a real to complex converter 22 which converts the digital samples to complex number values in order to provide a complex signal centred on zero frequency. This signal is fed to a Fast Fourier Transform device (FFT) 24 and to a timing synchronisation unit 26 which derives a value from the complex input signal which is fed to a digital to analog converter 28 to supply an analog control voltage to a voltage controlled oscillator 8, which provides a sampling clock signal to analog to digital converter 20.

The FFT device 24 has four modes of operation. Firstly, it is capable of performing either a 2048 point or an 8192 point transform. Second, it is capable of performing the transform in either direction. The inverse FFT functionality is provided so that the integrated circuit may be used in applications requiring OFDM modulation. In any event, the FFT performs a series of discrete Fourier transforms on each carrier frequency to provide at an output the data symbols for each carrier frequency. These output signals are corrected in phase at a common phase error generator unit 30 and then passed to a channel equaliser 32, a channel state information correction unit 34 and a deinterleaver 36. The signal thus processed is then passed at an output from the demodulator to forward error correction unit 8. The phase error correction block 30 calculates the common phase error of the signal and applies the necessary correction. The channel equaliser 32 first performs linear temporal equalisation followed by frequency equalisation using a high order interpolating filter. The equaliser outputs an equalised constellation to the channel state information unit 34. Unit 34 generates 3 or 4 bit soft decisions which are suitable for presentation to a Viterbi decoder. Deinterleaver 36 performs firstly symbol deinterleaving followed by bit deinterleaving.

In addition, the output signals from FFT 24 are passed to a frequency synchronisation unit 38 and converts it to a control signal for automatic frequency control, which acts upon a local oscillator in down-converter unit 2 for adjusting the frequency of the first or second IF.

In addition, the output of FFT 24 is fed to a frame synchronisation unit 40 which is fed forward to units 10, 12 and 14 of FIG. 1. A microcontroller interface 42 is provided, and in addition RAM memory 44 is provided to which all the units 22, 24, 30-36 have access to in order to provide their required operations.

Analog Versus Digital AFC

One of the processes that is required in the synchronisation of the demodulator is to obtain frequency synchronisation. There is a choice as to whether to apply the required frequency shift as an analog correction in the down-converter 2, or as a digital frequency shift in the demodulator chip.

Analog Frequency Correction

If the frequency correction is implemented by adjusting the frequency of the reference crystal in the down-converter 2, then a control signal on line 9 is provided from the output of the integrated circuit 6 back to the down-converter. This method has the advantage that a SAW filter inside the down-converter can be made as narrow as possible. The disadvantages are twofold. First, the integrated circuit must pass a control signal back to the down-converter. Second, the architecture of the down-converter is made more complicated since the control signal must adjust the reference crystal within the search range of the AFC.

Digital Frequency Correction

If the frequency correction is implemented in the integrated circuit 6, then the architecture of the down-converter 2 is made much simpler since there is no longer any need to have a control signal from the chip 6, and the loop in the down-converter that drives the reference crystal is no longer required. The disadvantage of this method is that the bandwidth of the SAW filter must be increased by the AFC search range. This causes a significant penalty in terms of the adjacent channel protection ratio when the receiver is used in an environment where the existing analogue services are operated in adjacent channels to digital services. The architecture described will permit both analog and digital correction.

As regards common phase error correction, this in practice combined with generation of the control signal in unit 38 for automatic frequency control. Both measurements are based on the phase rotation between one symbol and the next, measured on the continual pilots (CP's).

If a constant AFC error is present, there will be a constant change of rotation between successive symbols, proportional to the frequency error. Low frequency phase-noise will have a similar effect; rotating all of the carriers by the same angle, but this angle will vary from symbol to symbol in a random manner. In both cases it is desirable to attempt to correct the phase error on the current symbol by applying the opposite phase rotation to all carriers--this process is known as common-phase-error correction.

In addition to the phase rotation effect, an AFC error will also cause inter-carrier interference (ICI) which cannot be corrected for--for this reason it is also necessary to feed back an error signal to drive the frequency error to zero. This error signal can be applied to either in the analog domain as the local-oscillator control voltage, or in the digital domain to a DDFS which must be situated before the FFT. In either case an appropriate loop filter is included.

The measurement of phase rotations can only resolve AFC errors of up to roughly one half of the carrier spacing in either direction. In practice, during acquisition the AFC error is likely to be much greater than this. For this reason the AFC measurement also includes a "coarse" part, which measures the number of whole carriers by which the frequency is wrong. This is done using a pattern-matching approach looking for the continual pilots.

Referring to FIG. 4, a frequency offset can therefore be viewed as a shift of all the carriers either on the left of on the right. Basically, the frequency offset is divided in two parts.

1) Coarse Frequency Offset: A multiple of the carrier spacing

2) Fine Frequency Offset: A frequency offset less than the carrier spacing.

The OFDM signal is formed with a group of four different types of carriers, which are data carriers, continual pilots, scattered pilot and TPS pilots. Their positions are well defined by DVB-T specification. The continual pilots are always transmitted at the same position from OFDM symbol to OFDM symbol, as shown in FIG. 11, which is taken from the ETSI standard for DVB, no. ETS 300744; for each OFDM symbol, continual pilots transmit exactly the same known information.

A fixed frequency offset rotates all carriers with the same phase from symbol to symbol. Therefore, the first phase difference between two carriers at the same index k belonging to two consecutive OFDM symbol gives the amount of frequency offset modulo π. This can be shown as follows:

Symbol m, with N carriers, on a frequency F₀ lasting T_T with a carrier spacing of w_s may be written: ##EQU1##

The symbol is assumed to be integrated on Ts whereas it is sent through the Channel during T_T =T_s +T_Guard. Assuming a frequency offset of Δw₀ =nw_s +δ w₀, the l output of the FFT equals: ##EQU2##

Which gives for carrier 1 of symbol m ##EQU3##

Phase difference for each carrier between consecutive symbols: Φ=e-j(nw^.sub.s +δw^.sub.0 )T^.sub.T

Conclusions: All carriers are rotated with the same phase from symbol to symbol

As continual pilots always carry the same information, then this difference is constant with time. Therefore, the second difference (the difference of the difference) shall be zero for all continual pilots, and random values for all carriers (as data on these carriers are changing from symbol to symbol).

Applying the second phase difference, (the difference of the difference between two consecutive symbols at carrier index k) should lead to a null phase value whatever the frequency offset is. For all the carriers that are not continual pilots, this value shall be a random value.

Each OFDM symbol carries 2k or 8k carriers. Continual pilots (45 or 177) always lie at the same index and are spaced apart as indicated in FIG. 11 a variable distance, e.g. the sequence of carriers 0, 48, 54, 87, 141, 156 gives spacings of 47, 5, 32, 53, 14. Therefore, knowing the theoretical positions of the continual pilots, it is possible to search around the theoretical positions of the continual pilots to locate these zeros in second difference in phase. Further it is assumed frequency offsets are the same for all continual pilots in the symbol interval.

To achieve this, a bank of recursive filters is employed that accumulates the phase difference for each carrier around the theoretical position for a continual pilot signal, as indicated in FIG. 5. If a frequency offset of N carriers spacing (usually 47) is to be recovered, accumulator filters are placed around each continual pilot, N carriers before and N after. These accumulator filters contain the value of the second phase difference. Thus for each continual pilot CP1 . . . CPp, a single bank of filters is employed, each filter (k) accumulating the second phase difference for the carriers occupying the kth position within the search range N either side of the nominal carrier position. In order to filter the result for better performance, the values of each accumulator filter of index k (between -N and N) shall be added to give an average of the value for that position. Thus, a bank of accumulators contains the average value of the second phase difference, for each carrier of index k (k between -N and N) that lies at a distance of +or -N of every continual pilot. The index of the accumulator containing the smallest value is the coarse frequency offset.

In order to reduce the number of filters required, the values of the filters may be stored in a suitable storage means and accessed for each new data symbol interval. A problem arises because the continual pilot spacing is not constant, and if N is large (say 47), it is possible to have two or more than two continual pilots that lie in a range of 2N+1 (FIG. 11). If each continual pilot position were selected, then (e.g. k=48, 54, FIG. 11), then a carrier, e.g. at position 30 would be within the search range for each CP position. Requiring two simultaneous updates of the second phase difference. RAM memory cannot update at two simultaneous locations.

In accordance with the invention, a subset of the continual pilots is used. This subset shall have the property that no two consecutive pilots are spaced by less than 2N+1 pilots. Thus for example for a search range of ±47, the range 2N+1=97, and for k=255, the search will be carried out between 208 and 302. Thus in the 2K mode, the CPs selected are:

0, 141, 255, 432, 531, 636, 759, 873, 969, 1101, 1206, 1323, 1491, 1683. In addition to these, in the 8K mode, we have:

1791, 1896, 2037, 2136, 2235, 2340, 2463, 2577, 2673, 2805, 2910, 3027, 3195, 3387, 3495, 3600, 3741, 3840, 3939, 4044, 4167, 4281, 4377, 4509, 4614, 4731, 4899, 5091, 5199, 5304, 5445, 5544, 5643, 5748, 5871, 5985, 6081, 6213, 6318, 6435, 6603, 6795. Thus it will be seen the groups selected are non-overlapping and essentially contiguous.

It will be understood that in many of groups of carriers searched, there will be more than one CP present. Whilst the process of filtering over many groups will filter out a false selection made in one group, nevertheless problems remain as follows.

In COFDM channels for digital terrestrial TV, one might need to recover large frequency offset and/or be very robust against noise. In case of large frequency offset, say where N=47, and 2N+1=95, the cost is that fewer continual pilots can be assessed, since a large search window is required, which does not overlap with adjacent search windows, and many Continual Pilots (CP) may be present in the search window. This has two drawbacks: firstly immunity to noise decreases due to a lighter filtering. Secondly, as more than one CP may be present in the search window, the selection of the wrong local minimum may endanger the correct offset recovery, inducing false locks.

In the case of small SNR having more CP available for measurement would filter more the signal and thus increase the performance. Therefore what is required is a system with a good search range and a good noise immunity. In other words, a system capable of being locally precise and extendable.

In the present invention, as indicated in FIG. 7, a RAM memory is divided into three contiguous memory sections 700, designated 1, 0, 2. Each memory is a separate element in the chip being a first set of decoders 702 for selecting the minimum in each section 700, and a second decoder 704 for selecting the minimum value of the local minimum decoders 702. For each memory section 1, 0, 2, a corresponding filter F1, F0, F2 is provided to determine the second phase difference for a number of carriers P, corresponding to the number of memory locations 706 within each row. Each memory location 706 receives the result of the second phase difference for a respective frequency.

This method may be used in two different ways.

1) If the search range for the required frequency offset is too big for a single physical memory section, M blocks≦k≦N are used each with a maximum frequency offset of Δf. These M blocks are then capable of recovery of a frequency offset of MΔf. M is usually 2 or 3. Each block has the same number of continual pilots to average on, so there is no loss of robustness. This means M times more memories, but using higher clock rates and/or double port RAMs then M logical memories may be placed into one physical memory. This would lead to a smaller physical memory.

2) If the maximum recoverable frequency offset is sufficient, but the SNR is too small, instead of extending the search range, the memory is split into a set of smaller search ranges, as indicated in FIG. 7.

The big difference is that because there are M logical RAMs instead of 1 bigger RAM. Further such RAMs may operate simultaneously and independently of one another. The number of available continual pilots would increase as now all the CP with a distance greater then (2N+1)/M instead of 2N+1 can be selected. Each smaller RAM would then be more robust to noise because of a heavier filtering. Thus for N=47, 2N+1=95, and (2N+1)/M=32 (i.e. 16 carrier spacings, roughly, on either side of the CP). As before, these RAMs could be one physical RAM by either increasing the clock frequency. This method however does not increase the number of RAM cells.

The overall process is the following:

When a continual pilot is detected, the system enters the SWEEP mode, if it was not already SWEEPING. Then, for 2N+1 carriers, the system is going to accumulate the new value in the register of index k (k grows from -N to N). If the new value is smaller than the local minimum, then the minimum is updated to the register content, and the index of the minimum is set to k.

After 2N+1carriers, the mechanism stalls until the next continual pilot arrives. When all the continual pilots have been received, the index of the minimum contains the reliable index of the minimum, which in turn gives the coarse frequency offset.

Referring now back to the specific embodiment and to FIG. 8, this shows certain elements of FIG. 2 in more detail, in particular the common phase error circuit 30 and the channel equalisation circuit 32. In practice, CPE circuit 30 is combined with AFC circuit 38, and comprises a complex to phase format converter 50 for converting the symbol values output from FFT unit 50 to a phase format. These converted symbol values are fed into delay elements 52, 54, and subtract units 55, 56 derive the phase difference between the current signal and that stored in element 52, and the phase difference between the symbols stored in elements 52, 54. These phase differences are used to control phase error correction circuits 180, 182. In addition a subtractor 58 is used to determine the difference between the two signals provided by subtractors 55, 56. The output second difference signal is used for coarse frequency control, as indicated below. The symbol values stored in elements 52, 54, are fed via phase correction circuits 180, 182 to channel equalisation unit 32.

The incoming data are denoted by c(l,n) where l is the symbol number and n is the slot number within the symbol. Note that this is not the same as the carrier number k, because this block must start processing before the nominal position of the first carrier to allow for a coarse frequency error.

The incoming complex values are converted to phase: ##EQU4##

where the argument function is defined such that -π≦arg(z)<π. c(l,n) is also delayed by one and two symbols and converted to phase to give θ(l-1,n) and θ(l-2,n) in delay elements 52, 54.

The first difference of phase is calculated for the current and previous symbols in subtractor units 55, 56.

φ(l,n)=[φ(l,n)-φ(l-1,n)]mod1.0

φ(l-1,n)=[φ(l-1,n)-φ(l-2,n)]mod1.0

The second difference is also calculated in further subtractor unit 58.

ψ(l,n)=[(l,n)-φ(l-1,n)]mod1.0

The differences are calculated modulo 1.0, i.e. they are all between -0.5 and +0.5.

Referring to FIG. 9, one filter of the recursive filters F2, F0, F1 in FIG. 7 is shown. The magnitude of the second difference ψ(obtained from subtractor unit 58) is obtained in unit 82 and applied to filter 80 comprising subtractor 84, scaler 86, summer 88 and store 90, and a feedback loop 92 to summer 88 and subtractor 84. A unit 94 determines the smallest value in store 90. Store 90 and unit 94 are controlled by an offset counter 96. The value from counter 96 is provided as coarse frequency correction, corresponding to the smallest value in store 90.

Store 90 corresponds to a memory section 700 of the RAM of FIG. 7, and unit 94 corresponds to first decoder 702.

Thus the coarse AFC uses a bank of recursive filters having an output γΔ (l,n) in which each value of Δ corresponds to a different trial frequency offset. The search range is given by -47≦Δ≦47. Each filter is updated only when the current slot would contain a continual pilot for its particular value of Δ (see above). The input to all of the filters is the rectified value of the second difference of phase. This will have a small average value only for the correct offset, because the first difference will be similar each time. The update rule for the output of the filter is: ##EQU5##

where R is the scaling factor applied by scaler 86. Where C_R is a subset of C chosen such that at most one store needs to be updated for each slot as explained above. The store which needs to be updated, if any, is the one for which Δ=n-N₀- N_C, where N_CεC_R.

After all of the filter stores have been updated for a given CP, the coarse AFC output is set to the value of Δ corresponding to the store containing the smallest value:

Δ_C =argΔ minγΔ

The algorithm can be expressed as follows:

TBL if ((n-N₀ + 47) ε C_R && sweep == false) { Δ = -47; sweep=true; γ_min = 0.5; Δ_min = 0; } if (sweep) { γΔ = (1 - R)γΔ + R|Ψ|; if (γΔ < γ_min) { γ_min = γΔ; Δ_min = Δ; } Δ = Δ + 1; if (Δ>47) { sweep=false; Δ_C = Δ_min ; } }

where the flag sweep starts with the value false. The important value to be defined from this process is the offset value Δ_c, which represents the coarse frequency offset in terms of number of carrier spacings.

Two methods are used for combining coarse and fine measurement.

METHOD 1:

This method simply adds the two measures together, since the coarse measure is an integer giving the number of whole carriers offset, while the fine measure gives fractions of a carrier.

E=Δ_C +φ_W

METHOD 2:

In this method the fine part is only considered if the coarse value is zero. The coarse value is also clipped to the range -1 to +1: ##EQU6##

The frequency error value E is fed into a loop filter, which consists simply of an integrator:

v(l)=v(l-1)+E|_n=N_max

The integrator is clocked once per symbol, at the end of the symbol. This value is fed to a DAC which can be used to generate the AFC control voltage if analog AFC is being used. The value is also fed to the DDFS if digital AFC is being used.

Referring now to FIG. 10, this shows a circuit for combining the outputs of the fine and coarse error circuits . The input O5 is applied to an accumulator 100 which provides a common-phase-error signal. The signal O5 is also applied to a combining circuit 102 where it is combined with the output O6 from FIG. 9 in order to provide a summed signal which is applied to an accumulator 104. The output of accumulator 104 is applied to truncation circuits 106, 108. The output from truncation circuit 108 may be used for digital automatic frequency control where the signal is applied to a DDFS circuit at the input of FFT unit 24; alternatively the output from truncation circuit 106 is applied to a Σ/Δ digital to analog converter circuit 110 in order to provide a signal for analog automatic frequency control where the frequency of a local oscillator in the down-converter stage is controlled.

A major advantage of this method is that it requires no specific lock information, that would be difficult to derive in COFDM.

Besides, the duplication of identical hardware has the merit of simplicity and the choice in combination functions may lead to various degrees of rapidity/robustness.

Advantage of this method are as follows:

1) No lock indication required to recover the frequency offset using all the continual pilot. May lead to better performance, but on the other side, using a subset of the continual pilots lead to very optimised and cost effective solutions.

2) Easy extendible range. A simple duplication of these banks leads to increase of the maximum recoverable offset. Using specific error functions (with limiters) may lead to various degree of finesse.

3) Usage of RAMs instead of registers that reduces the hardware area.

4) No area overlapping in each process that leads to the use of low cost RAMs

Claims

1. An apparatus for demodulating digital video broadcast signals comprising data modulated on a multiplicity of spaced carrier frequencies, comprising:

analog-to-digital conversion means for providing a series of digital samples of the broadcast signal;

transform means for analyzing the samples to provide a series of data symbol values for each carrier frequency;

signal processing means for processing the series of data symbol values including phase error correcting means; and

automatic frequency control means for controlling the frequency of the signals input to the transform means;

wherein the automatic frequency control means includes coarse frequency control means for controlling the frequency in terms of increments of the carrier spacing frequency, and fine frequency control means for controlling the frequency for values less than a single carrier spacing frequency interval; and

wherein the coarse frequency means includes a filter means for assessing a group of a predetermined number (N) of carrier signals on either side of the nominal position of a plurality of predetermined continual pilot signals, wherein the outputs of the filter means are provided to respective memory locations of memory means, the memory means being divided into a plurality of sections, first means for determining within each section a first signal best representing the continual pilot signal, and second means for determining from among the first signals from the plurality of sections, a second signal which best represents the continual pilot signal.

2. An apparatus according to claim 1, wherein the plurality of sections comprises 2 or 3 sections.

3. An apparatus according to claim 2, wherein the filter means comprises one filter for each memory section.

4. An apparatus according to claim 1, wherein the filter means comprises one filter for each memory section.

5. An apparatus according to claim 1, wherein the filter means comprises a single filter.

6. An apparatus according to claim 5, including means for determining the rate of change of phase of the predetermined number of carrier signals between consecutive symbol intervals, and applying the same as inputs to respective filter means of the bank of filter means.

7. An apparatus according to any claim 1, including means for determining the rate of change of phase of said predetermined number of carrier signals between consecutive symbol intervals, and applying the same as inputs to respective filter means of the bank of filter means.

8. An apparatus according to claim 6, wherein the filter means comprises a recursive filter for providing an output representative of the accumulated value of a plurality of previous input values, said first means arranged for selecting the minimum output value as representing the continual pilot signal.

9. An apparatus according to claim 8, wherein each filter comprises a difference means coupled to the filter input, to provide an input signal to said memory locations, the output of the store means being fed back to the difference means, so that the output of the difference means represents the difference between the filter input and the output of the store means.

10. An apparatus according to claim 9, wherein the filter means includes an offset counter for shifting the address of the location within the memory means for storing the values at different locations.

11. An apparatus according to claim 1, wherein the filter means includes an offset counter for shifting the address of the location within the memory means for storing the values at different locations.

12. An apparatus according to claim 6, wherein said first means includes means for comparing a current filter output with a previously determined minimum output value, and updating the minimum value as necessary.

13. An apparatus according to claim 8, wherein said first means includes means for comparing a current filter output with a previously determined minimum output value, and updating the minimum value as necessary.

14. An apparatus according to claim 13, wherein only continual pilots are assessed which are spaced at least a predetermined number (2N+1) carriers apart.

15. An apparatus according to claim 1, wherein only continual pilots are assessed which are spaced at least a predetermined number (2N+1) carriers apart.

16. An apparatus according to claim 15, wherein only continual pilots are assessed which are spaced at least a predetermined number (2N+1)/M carriers apart, where M is the number of memory sections.

17. An apparatus according to claim 1, wherein only continual pilots are assessed which are spaced at least a predetermined number (2N+1)/M carriers apart, where M is the number of memory sections.

18. A method for demodulating digital video broadcast signals comprising data modulated on a multiplicity of spaced carrier frequencies, the method comprising:

analog-to-digital conversion to provide a series of digital samples of the broadcast signal;

analyzing the samples to provide a series of data symbol values for each carrier frequency;

processing the series of data symbol values including phase error correction and automatic frequency control of the frequency used to analyze the samples;

wherein the automatic frequency control includes coarse frequency control for controlling the frequency in terms of increments of the carrier spacing frequency and fine frequency control for controlling the frequency for values less than a single carrier spacing frequency interval;

wherein the coarse frequency control includes a filter means for assessing a group of a predetermined number (N) of carrier signals on either side of the nominal position of a plurality of predetermined continual pilot signals and providing outputs to respective memory location of the memory means, the memory means being divided into a plurality of sections;

determining within each section a first signal best representing the continual pilot signal; and

determining from among the first signals from the plurality of sections, a second signal which best represents the continual pilot signal.